Deep CABB observations of ECDFS Extended Chandra Deep
Deep CABB observations of ECDFS (Extended Chandra Deep Field South) Ray Norris, Jamie Stevens, et al.
Goals of the Observations: 1. Commissioning • Are there weak artefacts visible only in a deep image? – Spike at phase centre – Stripes from internal interference – Closure errors causing calibration problems • Will we be able to get to thermal noise limit? – Could be limited by artefacts • Need to observe a field free from strong sources
Goals of the Observations: 2. Astrophysics • Observations part of scheduled ATLAS project C 2028 for deep 6 cm observations of ECDFS – ECDFS has the deepest optical, infrared and X-Ray imaging available anywhere (and more to come with Herschel etc) – Goal: 8 µJy rms - the widest 6 cm survey of this sensitivity. • Try to solve link between AGN and starbursts in the faint radio population. • E. g. – Measure spectral indices – probe weak GPS (“baby radio-loud AGN”) – Look at morphology of AGNS buried inside SF galaxies
A puzzle • S 442 & S 443 are two interacting star -forming galaxies at z=0. 076 (both detected with Chandra) • C 473 & C 476 were previously identified as two SF galaxies. • We think they are two lobes of a radio-loud AGN (P=3. 4 × 1023 W/Hz). These are almost always associated with elliptical galaxies. • BUT the source between them (z=0. 214) is a Sab galaxy! • What gives? 20 cm ATCA contours (rms ~ 20 µJy) overlaid on 3. 6µm Spitzer image. Project C 1950 (Mao et al. ) observed this source at 6 cm to measure α, but failed to detect the source, possibly because of the 6 km configuration they used. Can we go deeper and measure α?
Observations • ECDFS observed by Jamie Stevens on 19 April for ~7 h • After removing bad data (source setting), calibration, etc, get ~ 4. 1 h on-source. • Calibration above 6. 2 GHz unreliable • Some channels missing due to bad correlator boards • Some interference spikes • Otherwise looks good Calibrator – before editing ECDFS – after editing
First pass of imaging: simply add the 2 GHz bandwidth together and process using Miriad First image using mfs, mfclean, but no selfcal. Rms=21µJy. Theoretical rms ~14µJy. No visible artefacts, even at phase centre.
As before, but using (phase) selfcal Rms=16µJy. Theoretical rms ~14µJy. Selfcal has reduced the rms, but has introduced some calibration artefacts, presumably because of differing spectral indices. But still no artefact at phase centre.
20 cm contours (6 km array) overlaid on 6 cm greyscale (H 168 array). This source previously undetected with 6 cm 6 km array. Now 1 m. Jy of flux visible on H 168 array. S(20 cm) = 5. 1 m. Jy so α=-1. 2 DEFINITELY AN AGN!!!
Conclusions • Even with channels missing etc, CABB is b****y fantastic! • In 4 h, we got deeper (16µJy) than we could have got in 12 h (20µJy) using the old correlator. • Science result: we detected a 6 cm source confirming a Radioloud AGN buried inside an Sab galaxy (almost unknown in the local Universe). • No obvious artefacts other than the known interference spikes. • Using a brute-force analysis method gets close to thermal noise limit in this field which is dominated by faint sources (possibly with a fairly uniform spectral index). • Using a more careful technique of splitting into bands has not yet given a significant improvement (but will probably be essential on more complex sources).
- Slides: 9